The importance of developing a predictive toxicity paradigm to assess ENM hazard in the lung Pulmonary toxicity as a result of inhaling engineered nanomaterials (ENM) depends on the unique physicochemical properties that allow these materials to perturb bio-molecules and bio-molecular processes in the lung.{1} We define the nano-bio interface as the interacfion of ENM surfaces, which are shaped by intrinsic material properties as well as the dynamic modificafion of those properties by environmental media, with proteins, DNA, membranes, lipids, cell surfaces, endocytic pathways, intracellular organelles, cytosol, nucleus, biological fluids, fissue and organs.{2} {1}While ENM-based products such as nanocomposites, surface coafings and electronic circuits are unlikely to pose a direct risk to the lung, ENM that are being produced as nanoparticles, agglomerates of nanoparticles or particles comprised of nanostructured materials are more likely to pose a hazard to the lung.^ While it is theorefically possible to subject every new material that is being produced as an unattached particle to rigorous inhalafion toxicity testing in animals, this is logisfically unfeasible at the rates at whicti new ENM are being produced, including cost and animal use considerafions. This limits the number of different material composifions that can be studied in animals as well as the ability to assess all the physicochemical properties that can be engineered into one material, including size, surface area, shape, crystallinity, surface charge, reactive surface groups, dissolufion, state of aggregation or dispersal etc. It is our opinion that knowledge generafion about ENM hazard has to consider additional approaches that complement animal testing.{3} In this proposal, we recommend the implementation of a predictive toxicological paradigm, which is defined as the assessment of in vivo toxic potential of ENM based on in vitro and in silico methods.{3} Predictive toxicology is an essential tool for successful drug development because toxicity is one of the major reasons for product failure in the drug development process. It is essential to identify and exclude new drug candidates with unfavorable safety profiles as early as possible in the development process. Predictive toxicology has recently also being introduced to industrial chemical toxicity. Both the Nafional Toxicology Program as well as the Nafional Research Council (NRC) in the US Nafional Academy of Sciences (NAS) have recommended that toxicological testing in the 21st-century evolve from a predominanfiy observafional science at the level of disease-specific models to predictive science models focused on broad inclusion of target-specific, mechanism-based biological observations.{4-6} It is further recommended that the biological testing be based on robust scientific paradigms that can be used to screen mulfiple toxicants at one fime instead of costly animal experiments looking at a single toxicant at one fime. A report outlining the US Federal Government response to the NRC document was published in 2008 and prompted NIEHS, EPA and the National Institute of Health Chemical Genomics Center to sign an agreement to collaborate on the development and evaluation of a rapid and high volume screening methodologies to: (i) prioritize substances for more comprehensive toxicological tesfing, (ii) identify mechanisms of acfion for further invesfigafion, and (iii) develop predictive models for in vivo biological response monitoring for commercial chemicals with inadequate or nonexistent toxicological data. Although this change in toxicological assessment philosophy has catalyzed a healthy and rigorous debate among toxicologists, regulators and the public, our opinion is that it is fimely to consider an analogous approach for ENM hazard assessment. Importanfiy, we do not recommend doing away with animal experiments but we advocate the use of toxicological or mechanistic injury pathways to establish in vitro property-activity relafionships that can be used for knowledge generafion and logical planning of animal testing. Project 2 will determine whether the property-activity relafionships to be explored by carefully chosen and wellcharacterized compositional and combinatorial ENM libraries can help us understand the material properties leading to pulmonary inflammation, cytotoxicity and fibrosis. Integral to understanding these properties is the ability to develop dosimetry models that consider biological hazard in dose quantifies other than mass.{2}